<div dir="ltr"><div class="gmail_quote"><div dir="ltr">On Wed, 31 Aug 2016 at 21:54 Henri Sivonen <<a href="mailto:hsivonen@mozilla.com">hsivonen@mozilla.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">> No work is done on stable.<br>
<br>
This looks like a policy issue to me. Are there technical blockers<br>
that the paragraph above wouldn't address?<br></blockquote><div><br></div>Perhaps we should rewind a long way and explain why we make releases at all.</div><div class="gmail_quote"><br></div><div class="gmail_quote">"Releases" of Debian are self-consistent snapshots of the ongoing work of Debian maintainers. Publishing a release serves a few purposes in my mind:</div><div class="gmail_quote"><br></div><div class="gmail_quote">- It gives a fixed version of all the packages to receive additional integration testing, long-term support statements, and (later) security updates</div><div class="gmail_quote">- It gives an easy label for 3rd parties to point at and say "My software works with *that* version"</div><div class="gmail_quote">- It allows fixed/offline media (DVDs, etc) to be published with a known set of self-consistent contents</div><div class="gmail_quote">- It declares an epoch beyond which Debian maintainers no longer have to support package upgrades<br><div><br></div><div>Yes, we could invent some new concept where we do all the above .. and then we continue to update packages wherever someone decides that newer is better. Such a "rolling release" fails to meet any of the above motivations and is roughly the same as not declaring a release at all - and indeed this continually updating thing is pretty much exactly what the Debian "testing" and "unstable" "rolling releases" are. To phrase it another way: If we only had rolling releases, we would *still* have strong reasons to invent a fixed-point-in-time release concept in order to satisfy the above purposes. *Not changing* is the *entire point* of a stable release.</div><div><br></div><div>Now as a project, we've recently decided that browsers are sufficiently difficult to maintain security for that we're going to try to make exceptions for them in this "snapshot" concept. I think you can see from the various public statements mentioning the browsers by name, and the lack of consistent well-understood policy just how exceptional these are. Following the "what if everyone did this?" principle, I hope you can see that we *really don't* want to extend these exceptions and start making changes to other parts of past stable releases, even though copying some non-conflicting packages into the archive is technically feasible (and in many cases, easy).</div><div><br></div><div>I hope this makes it a little clearer...</div><div><br></div><div> - Gus</div></div></div>